Chemosphere 68 (2007) 18541860

www.elsevier.com/locate/chemosphere

Enhanced ozonation of simulated dyestu wastewater by microbubbles

Li-Bing Chu a, Xin-Hui Xing
a

a,*

, An-Feng Yu a, Yu-Nan Zhou a, Xu-Lin Sun b,

Benjamin Jurcik b

Department of Chemical Engineering, Tsinghua University, Beijing 100084, China

b
Air Liquide Laboratories, Tsukuba-shi, Ibaraki-Pref. 300-4247, Japan

Received 20 December 2006; received in revised form 9 March 2007; accepted 9 March 2007
Available online 27 April 2007

Abstract
The ozonation of synthetic wastewater containing azo dye, CI Reactive Black 5, was investigated using a microbubble generator and a
conventional bubble contactor. The microbubble generator produced a milky and high intensity microbubble solution in which the bubbles had a mean diameter of less than 58 lm and a numerical density of more than 2.9 104 counts ml1 at a gas ow rate of less than
0.5 l min1. Compared with the bubble contactor, the total mass transfer coecient was 1.8 times higher and the pseudo-rst order rate
constant was 3.23.6 times higher at the same initial dye concentration of 100 mg l1, 230 mg l1 and 530 mg l1 in the proposed microbubble system. The amount of total organic carbon removed per g of ozone consumed was about 1.3 times higher in the microbubble
system than in the bubble contactor. The test using terephthalic acid as the chemical probe implied that more hydroxyl radicals were
produced in the microbubble system, which contributed to the degradation of the dye molecules. The results suggested that in addition
to the enhancement of mass transfer, microbubbles, which had higher inner pressure, could accelerate the formation of hydroxyl radicals
and hence improve the oxidation of dye molecules.
 2007 Elsevier Ltd. All rights reserved.
Keywords: Microbubbles; Ozonation; Mass transfer; Hydroxyl radical; Dyestu wastewater

1. Introduction
Ozone is a well-known powerful oxidant. It is widely
used to improve the taste and color of drinking water as
well as to remove organics and inorganic compounds in
drinking water and wastewater (Kasprzyk-Hordern et al.,
2003). The ozonation process is a typical example of gas
absorption with a chemical reaction where the total reaction rate can be aected by both the reaction kinetics and
mass transfer (Zhou and Smith, 2000). It has been determined that the rate-limiting step comes from the low gas
liquid mass transfer rate due to the low solubility of ozone
(Danckwerts, 1970). The rate of ozone mass transfer
depends on the mixing characteristics of the gasliquid

Corresponding author. Tel./fax: +86 10 62794771.

E-mail address: xhxing@tsinghua.edu.cn (X.-H. Xing).

0045-6535/$ - see front matter  2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chemosphere.2007.03.014

contactor used, the kinetics of ozone decay in the water,

and the number and size of the ozone bubbles produced
(Shin et al., 1999; Rosal et al., 2006). The eectiveness of
ozonation can be increased by a higher surface area of
ozone through the generation of smaller bubbles (Mitani
et al., 2005).
Many factors hinder the use of ozone, including its relatively low solubility and stability in water, the high cost of
ozone production, and the selectivity of oxidation. Ozone
can slowly react with some organic compounds such as
aromatics (Kasprzyk-Hordern et al., 2003). Moreover, in
many cases, ozonation cannot completely mineralize the
organics to carbon dioxide, leading to partial oxidation
products such as aldehydes, organic acids and ketones
(Slavinskaya and Selemenev, 2003). In order to enhance
ozonation eciency, much attention has been paid to the
development of advanced oxidation processes using O3/
H2O2, UV/O3, or ultrasound/O3, and catalytic ozonation

L.-B. Chu et al. / Chemosphere 68 (2007) 18541860

using metal ions, metal oxides or activated carbon

(Ruppert et al., 1994; Kasprzyk-Hordern et al., 2003).
These processes mainly aim at the generation of hydroxyl
radicals, which are more powerful oxidant than molecular
ozone. Up to now, these advanced techniques are mainly in
the laboratory investigation stage due to the complexity
and high cost of these techniques. To make full use of
the advantages of ozonation in industry, further study is
needed to nd ecient techniques to enhance the mass
transfer and oxidation of ozone.
Microbubbles are dened as bubbles with diameters less
than several tens of lm. Compared to conventional bubbles with diameters of several mm, microbubbles have a
huge interfacial area and bubble density, low rising velocity
in liquid phase and high inner pressure (Serizawa et al.,
2003). Microbubbles have shown promise due to their
potential to be widely used in such elds as biomedical
engineering, environmental engineering, drug reduction in
shipbuilding, marine culture, mining and domestic bathes
(Burns et al., 1997; Kodama et al., 2000; Soetanto and
Chan, 2000; Walker et al., 2001; Serizawa et al., 2003).
The application of microbubble technology in ozonation
processes provides a new approach for improving ozonation eciency.
In this paper, a widely used azo reactive dye, CI Reactive Black 5 (RB 5) was chosen as a representative model
dye to study the performance of ozonation by microbubble
technology. Azo dyes are non-biodegradable by the conventional activated sludge process, which causes serious
disposal problems (Vandevivere et al., 1998). Ozonation
has been reported to be very eective in reducing the color
of dyestu wastewater, but it is not eective in mineralization of organic substances (Hao et al., 2000; Wang et al.,
2003). The main purpose of our study was to examine
the enhanced mass transfer and the ozone oxidation of dyestu wastewater by microbubble technology. The mechanism of enhanced degradation of the dye by ozone
microbubbles was also analyzed.

2. Materials and methods

2.1. Experimental apparatus
The microbubble generator was made by Kyowa Engineering Co., Ltd. (Japan). It consisted of a recycling pump,
a gyratory accelerator and an injector (Fig. 1). The principle of microbubble generation is described in the Supplementary Data. The working volume of the Plexiglas
reactor was 20 l. Ozone was produced from pure oxygen
by an ozone generator (ED-OG-R4, Eco Design, Japan)
and fed into the reactor by the microbubble system or a
conventional gas diuser, which was used for comparison.
The titanic porous diuser was cylindrical, with a pore size
of 45 lm. A 2% (w/v) potassium iodine solution was used
to adsorb the ozone remaining in the o gas at the exit
of the reactor.

1855

Fig. 1. Schematic diagram of the experimental set-up.

2.2. Materials
A commercial dye, RB 5, with the chemical formula of
C26H21N5O19S6 4Na (Scheme 1) and molecular weight of
991.8, was purchased from Tianjin Dyestu Chemical Factory (China) and used without further purication. The
synthetic wastewater was prepared by dissolving the RB
5 in deionized water.

2.3. Analysis
The concentration of ozone in the gas phase was analyzed by the iodometric method with KI solution (Pires
and Carvalho, 1998). The liquid-phase ozone concentration was measured using the indigo colorimetric methods
(Bader and Hoigne, 1982). The RB 5 concentration and
decolorization capacity during the ozonation process were
determined from the absorption intensity at its maximum
visible absorption wavelength (595 nm) by a 722s spectrophotometer (Shanghai Precision & Scientic Instrument
Co., Ltd., China). The optical spectra of a sample in the
UVvis adsorption wavelength (200800 nm) were scanned
by a UVvis spectrometer (UV 757, Shanghai Cany Precision Instrument Co., Ltd., China). The total organic carbon (TOC) concentration was measured using a TOC
analyzer (TOC-5000A, Shimadzu Corporation, Japan).
Anions were analyzed using a high performance liquid
chromatograph (10A, Shimadzu Corporation, Japan)
OH

NH2

N=N

N=N

NaO3S

SO3Na

M:-SO2CH2OSO3Na
Scheme 1. Chemical structure of RB 5.

1856

L.-B. Chu et al. / Chemosphere 68 (2007) 18541860

equipped with a shim-pack IC-A3 column and a conductivity detector (CDD-10AVP, Shimadzu Corp., Japan). The
mobile phase was 8 mM p-hydroxybenzoic acid, 3.2 mM
bistris and 50 mM boric acid. The pressure and temperature of the column were 6.86 MPa and 40 C,
respectively.
The concentration of hydroxyl radicals produced from
ozone was tested qualitatively using terephthalic acid
(TA) as the chemical probe. This method has been used
to detect hydroxyl radicals produced in aqueous phase
pulsed electrical discharge reactors and by ultrasound
(Fang et al., 1996; Mayank and Locke, 2006). Briey, TA
reacts with the hydroxyl radicals to form 2-hydroxytephthalic acid, which produces bright stable uorescence when
the excitation wavelength is set at 315 nm and the emission
is measured at 425 nm. In the present study, ozone was
absorbed into deionized water(R = 18.2 MXcm) at a temperature of 7 2 C and pH of 2.0 0.2. 5 ml samples
were withdrawn from the reactor to react with 5 ml of
0.5 mM TA. Then the samples were analyzed with a uorescence spectrophotometer (F-2500, Hitachi, Japan).
The size of the microbubbles was measured using a
microscope (E600, Nikon, Japan) equipped with a CCD
digital camera (SPOT RT SUPER, Diagnostic Instruments, Inc., USA). The recorded images were analyzed to
determine the bubble sizes using Image-Pro Plus software
(Media Cybernetics Incorporated, USA). After the produced microbubbles reached stability, 1 cm-long cells were
lled with samples, and immediately closed. The preparation and observation were conducted immediately to
ensure that there was no obvious change in the microbubble size inside the fully lled and closed cells. This method
might not accurately reect the distribution of the original
bubble sizes because some large bubbles may have been
broken during the sampling, but it did provide a rough idea
of the features of the microbubble generator.
2.4. Experimental procedure
The experiments were divided into two parts. First, the
enhancement of the mass transfer was studied. The temperature of the reactor was kept at 18 2 C. The ozone gas
was absorbed into deionized water until the ozone in the
reactor reached saturation. Because the self-decomposition
constant (kd) of ozone is much lower than the total mass
transfer coecient(kLa), the inuence of kd was not taken
into account in the derivation of kLa. Consequently, the
values of kLa could be determined by the following
equation:
dC
k L aC   C
dt

where C is the concentration of dissolved ozone in the bulk

liquid phase (mg l1) and C* is the concentration of saturated ozone in water (mg l1). The gas holdup (eg) was calculated by measuring the liquid height in the reactor with
or without gas dispersion.

Secondly, the ozonation of the simulated dyestu wastewater was studied. Prior to this experiment, a three-wayvalve was used to bypass the ozone-containing gas into
the KI solution, with a gas ow rate of 0.5 l min1, to
determine the inlet ozone concentration and to ensure the
stability of the gas. The input ozone concentration was
approximately 132 mg l1. Then the ozonation process
was initiated when the ozone-containing gas was sparged
into the reactor, which was operated in semi-batch mode
by feeding the ozone continuously. The temperature was
kept at 20 2 C. Samples were taken at appropriate time
intervals to analyze the color, pH and TOC. In addition,
the ozone concentration in the o-gas was detected during
the ozonation. The ozonation experiments were performed
using both the microbubble generator and a conventional
bubble contactor for comparison. All the experiments were
repeated to ensure precision and the average value for each
parameter was used. The relative standard deviation was
less than 6% for most of the measured data.
3. Results and discussion
3.1. Enhanced mass transfer of ozone by microbubbles
With the introduction of gas, the water in the reactor
gradually became almost milky because many ne bubbles
were formed. After gas feeding was stopped, the water
remained milky and bubbly for 45 min, indicating the very
low upward velocity of microbubbles. The eg depended on
the gas ow rate (see Supplementary Data). As expected,
the eg in the microbubble system was higher than that in
the bubble diused system due to both the high surface
area-to-volume ratio and the increased bubble density. It
was observed that the lower the gas ow rates, the smaller
the bubbles were produced in the microbubble generator.
Therefore, the dierence of the eg between the two systems
became smaller with the increase of the ow rate. A mixture of microbubbles and larger bubbles (approximately a
few mm in diameter) was obtained when the gas ow rate
was higher than 0.5 l min1.
By observing more than 600 bubbles, the bubble size distribution in terms of count as a function of gas ow rates
was obtained (see Supplementary Data). The count mean
diameter, the total number of gas bubbles per volume
and the specic surface area were calculated according to
the equations described in Supplementary Data. A mean
diameter of less than 58 lm, numerical density of more
than 2.9 104 counts ml1, and specic surface area of
more than 334 m1 were obtained at a gas ow rate of less
than 0.5 l min1. With the increase of the gas ow rate,
more bubbles with larger sizes were observed. Accordingly,
the bubble density and the surface area-to-volume ratio
decreased. The obtained high bubble concentration and
large specic surface area demonstrated the high intensity
of microbubble generation.
Fig. 2 shows changes in the concentrations of dissolved
ozone and o-gas ozone with time in the two systems at a

L.-B. Chu et al. / Chemosphere 68 (2007) 18541860

90

16

80
70

12

60

10
8

50

-1

kLa=0.143 min

40

30
4
20
2

-1

14

Output ozone conc. (mg l )

-1

Dissolved ozone conc. (mg l )

-1

kLa=0.256 min

10

0
0

10

20

30

40

50

Time (min)
Fig. 2. Time course of dissolved and output ozone concentrations in
deionized water at a gas ow rate of 0.5 l min1. Symbols: circles
microbubble generator; triangles bubble contactor.

gas ow rate of 0.5 l min1. The concentrations of dissolved ozone rose rapidly in the initial period because of
the large driving forces for the mass transfer of ozone
and then reached saturated. Clearly, the microbubble generator exhibited more ecient mass transfer of ozone and
higher concentrations of dissolved ozone compared to the
bubble contactor. The total mass transfer coecient in
the former was 1.8 times higher than that in the latter. This
is consistent with the expectation that smaller bubbles will
result in higher rates of mass transfer of ozone from the
bubbles by maximizing the surface area-to-volume ratio.
3.2. Enhanced ozonation of dyestu wastewater by
microbubbles
As the ozonation progressed, the color of the RB 5
solution was gradually bleached from black, to brown, to
yellow, and then to colorless. Fig. 3 presents the results

1857

of the ozonation of the RB 5 solution at various initial

dye concentrations. The results showed that the lower the
initial dye concentration, the higher the decolorization rate.
This dependency of decolorization rate on the initial dye
concentration may be due to the production of more than
one kind of intermediate during the ozonation processes
(Shu and Huang, 1995). The reaction kinetics could be
treated as a pseudo-rst order reaction and the pseudo-rst
order rate constant k was used to evaluate the decolorization eciency. The same pseudo-rst order behaviors have
also been reported elsewhere (Shu and Huang, 1995; Wu
and Wang, 2001). The pseudo-rst order rate constant in
the proposed microbubble system was 3.23.6 times faster
than that in the bubble contactor at the same initial dye
concentration. Considering the fact that the total mass
transfer coecient in the former is 1.8 times higher than
that in the latter, we suggested that microbubbles might
aect the degree of ozone oxidation in addition to enhancing the mass transfer.
The UV absorption spectra at dierent ozonation time
were obtained to study the destruction of the dye molecules. RB 5 exhibited absorption bands at 595 and
311 nm. The disappearance of the visible band was due
to the fragmentation of the azo link due to ozone molecule
and hydroxyl radical attack, which has been proposed as
the rst step in the degradation of azo dyes (Ince and Tezcanli, 2001). In addition to this rapid bleaching eect, the
decay of the absorbance at 311 nm was considered as evidence of the degradation of aromatic groups in the dye
molecules and their intermediates. The abatement rates at
595 nm and 311 nm are shown in Fig. 4. Degradation of
aromatic dye moieties proved to be a more recalcitrant
process. Ozone could cleave the conjugated double bonds
more thoroughly. The benzene rings could also be
destroyed to some extent. Clearly, the microbubble system
exhibited a faster abatement rate than the bubble
contactor.

1.0

1.0

530 mg l
-1
230 mg l
-1
100 mg l

0.8

C/Co

Normalized absorbency (At /A0 )

-1

0.6
0.4
0.2
0.0

Symbols: hollow-311 nm
solid-595 nm

0.8

0.6

0.4

0.2

0.0

20

40

60

80

100

120

140

160

Time (min)
Fig. 3. Degradation of RB 5 solution by ozonation at a gas ow rate of
0.5 l min1. Symbols: circles microbubble generator; triangles bubble
contactor.

20

40

60

80

100

120

Time (min)
Fig. 4. Rate of abatement at two absorption bands at an initial
concentration of 230 mg l1 during the ozonation of RB 5 solution.
Symbols: circles microbubble generator; triangles bubble contactor.

1858

L.-B. Chu et al. / Chemosphere 68 (2007) 18541860

3.3. Analysis of the enhancement mechanism by

microbubbles
To further examine the mechanism of the enhanced
ozonation by microbubbles, the performance of the two
systems was compared in terms of decolorization, TOC
removal, nitrate and sulfate concentrations, ozone utilization rate and enhancement factor due to chemical reaction
E at an initial dye concentration of 230 mg l1 (see Fig. 5
and Supplementary Data). The E value is dened as the
ratio of the rate of absorption with chemical reaction to
the maximum rate of physical absorption. It was calculated
according to the method by Zhou and Smith (1997). As
demonstrated in Fig. 5, the color removal initially occurred
rapidly, and slowed down, nally reaching 100%. The rate
of decolorization was obviously faster in the microbubble
system. The time required for 99% removal of color was
about 30 min after introduction of ozone into the microbubble generator, compared to 70 min for the bubble
contactor.
In terms of mineralization of the RB 5 solution, the TOC
removal eciency could be increased markedly by using

Color removal efficiency (%)

80

80

60

60
40

40

20

20

TOC removal efficiency (%)

100

100

0
0

20

40

60

80

100

120

100

Enhancement factor E

6
5

80

60

3
40
2
20

1
0

0
0

Ratio of consumed to applied ozone (%)

Time (min)

Applied ozone dose (g)

Fig. 5. Performance of the two systems in color, TOC removal eciency,
ozone utilization rate and E at an initial concentration of 230 mg l1.
Symbols: circles microbubble generator; triangles bubble contactor.

microbubble technology. The TOC removal eciency of

the microbubble generator reached 80% (at 80 min); it
was only 34% (at 130 min) for the bubble contactor. It is
interesting to note that the TOC removal lagged behind
the color removal during the ozonation. As shown in
Fig. 5, when almost 95% of the color had been removed,
the TOC of the solution began to decline gradually. This
could have been because the ozone rst attacked the chromophore, the azo link groups of the RB 5, and produced
intermediates (Ince and Tezcanli, 2001). Then, the reaction
between the ozone and the intermediates containing organic
carbons became dominant, which resulted in TOC reduction.
The sulfate and nitrate concentrations were measured
during the ozone treatment of RB 5 wastewater by ozone
(see Supplementary Data). The formation of nitrate was
most likely the result of the oxidation and cleavage of the
azo and amino groups (Wang et al., 2003). The cleavage
and oxidation of the sulfonic acid resulted in the increase
of sulfate. The high initial sulfate concentration of the
aqueous RB 5 solution was presumably caused by the presence of sulfate salts within the dye. The concentrations of
sulfate released and nitrate formed were higher for the
microbubble generator than those for the bubble contactor. These results were consistent with the TOC reduction.
The concentration of o-gas ozone of the microbubble
system (0.042.5 mg l1) was signicantly lower than that
for the bubble contactor (0.0874.3 mg l1), which demonstrated the very high ozone utilization of the microbubble
system. The ratio of the consumed ozone to the applied
ozone was used as a quantitative indicator to represent
the ozone utilization rate, as depicted in Fig. 5. The applied
ozone MA (mg) and consumed ozone MR (mg) are dened
by Eqs. (2) and (3), respectively:
M A QG  C A0  t

MR = Ozone applied ozone exited ozone accumulated

in holdup gas and liquid
Z t

QG C A0  C Ae dt  C Ai V H  C Ab V L
3
0

where QG is the gas ow rate (l min1), CA0 is the input

ozone concentration (mg l1), t is the reaction time (min),
CAi is the ozone concentration in the holdup gas (mg l1),
and is taken as the o-gas ozone concentration CAe in the
calculation for simplicity, VH is the volume of holdup gas
(L), VL is the working volume of the reactor (L). The dissolved ozone concentration (CAb) was found to be lower
than 0.08 mg l1 after the completion of decolorization,
so it was not included in this calculation.
When the bubble contactor was used, the MR/MA value
decreased slowly, then sharply. The decreasing of the MR/
MA value might have been caused by the lower reactivity of
the dye solution with ozone. It was thus proposed that the
initial decolorization reaction occurred readily, and the
rate of both the reaction and ozone utilization was high.
This also was conrmed by the signicant enhancement

Fluorescence intensity

2000

TA+ micro-ozone-bubble water

1600

TA+normal-ozone-bubble water

1859

0.20

-1

of the mass transfer, which was primarily due to the fast

ozonation reaction between the dissolved ozone and RB
5 solution (Qiu et al., 2001). At the later stage, the ozonation products were hard to oxidize, resulting in a much
lower reaction rate. The E value reached unity, indicating
the chemical reaction had little inuence on the mass transfer rate. Chen et al. (2005) investigated the ozonation of
RB 5 by using a rotating packed bed (RPB) and completely
stirred tank reactor (CSTR) as ozone contactors. The
experimental results indicated that the MR/MA value
decreased with ozone doses from 0.52 to 0.14 in the RPB
and 0.80 to 0.18 in the CSTR.
However, by using the microbubble generator, the MR/
MA value and enhancement factor decreased a little for
the applied ozone doses as ozonation proceeded, indicating
high reactivity during the whole ozonation process. Considering the high TOC removal eciency in the microbubble
system described above, we propose that the microbubble
technology can enhance the ozone oxidation as well as the
mass transfer. In general, ozone oxidation pathways include
direct ozone attack or free radical attacks. Since the oxidation potential of hydroxyl radicals is much higher than that
of ozone molecules, direct oxidation is slower than radical
oxidation. Microbubbles with higher internal pressure
could induce the decomposition of ozone and contribute
to the production of hydroxyl radicals. According to the
YoungLaplace equation DP 2cr (DP is the pressure dierence, c is the surface tension, r is the radius of bubble), the
smaller the bubble size, the higher the internal pressure at
this conditions.
This suggestion was proven by hydroxyl radical detection (see Fig. 6). The uorescence intensity of the samples
in the two systems increased with the continuous ozone
injection and then reached steady state (see Supplementary
Data). The uorescence intensity of the samples from the
microbubble system was much higher than that from the
bubble contactor. Ozone molecules could not react
with TA. There was almost no uorescence detected in

Removed TOC/consumed ozone (gg )

L.-B. Chu et al. / Chemosphere 68 (2007) 18541860

0.15

0.10

Microbubble generator
Bubble contactor

0.05

0.00
1

Applied ozone dose (g)

Fig. 7. Ratio of removed TOC value (g) to consumed ozone dose (g) in the
two systems at an initial concentration of 230 mg l1.

the presence of the hydroxyl radical scavenger of 90 mM

T-butanol in the ozone water sample. These results implied
that more hydroxyl radicals were produced in the microbubble system, which contributed to the degradation of
the dye molecules.
In order to eliminate the eects of the mass transfer, the
ratio of the removed TOC (g) to the consumed ozone dose
(g) was plotted as a function of the implied ozone dose (see
Fig. 7). This ratio is the amount of removed TOC per g of
consumed ozone. It was calculated when the TOC of the
solution began to decrease steadily, which was from
25 min to 90 min for the microbubble system and from
50 min to 130 min for the bubble contactor. It was obvious
that for the same ozone consumption, the TOC removal in
the former was 1.3 times higher than that in the latter indicating its higher reactivity.
As shown above, the microbubble technology will be a
promising process for the enhanced ozonation of dyestu
wastewater. One of the important factors pertaining to
the industrial application is the energy cost. Compared to
the conventional ozonation process, a liquid pump will
be required for the microbubble system, but the consumed
ozone doses were greatly decreased because of its higher
reaction activity. The total energy cost will be further analyzed by optimizing the scale and operation of the system.
4. Conclusions

1200
800
ozone water+T-butanol+TA

400
TA

ozone water

0
350

400

450

500

550

600

650

A microbubble technology was applied to increase the

mass transfer rate of ozone as well as to enhance the ozone
oxidation of synthetic RB 5 wastewater in this paper.
Experiments were performed using both the microbubbles
generator and the bubble contactor, which is commonly
used in ozonation systems, for the purpose of comparison.
Main conclusions obtained are as follows:

Wavelength (nm)
Fig. 6. Fluorescence detection of hydroxyl radicals formed in ozone water
under dierent conditions.

(1) The microbubble generator produced a milky and

high intensity microbubble solution in which the bub-

1860

L.-B. Chu et al. / Chemosphere 68 (2007) 18541860

bles had a mean diameter of less than 58 lm, a

numerical density of more than 2.9 104 counts ml1,
and a specic surface area of more than 334 m1 at a
gas ow rate of less than 0.5 l min1.
(2) Compared with the bubble contactor, the total mass
transfer coecient was 1.8 times higher and the
pseudo-rst order rate constant was 3.23.6 times faster at the same initial dye concentration in the proposed microbubble system at a gas ow rate of
0.5 l min1.
(3) The amount of TOC removed per g of ozone consumed was about 1.3 times higher in the microbubble
system than in the bubble contactor. The results of
the test using TA as the chemical probe implied that
more hydroxyl radicals were produced in the microbubble system, which contributed to the degradation
of the dye molecules.
(4) According to the results of mass transfer, decolorization, mineralization and ozone accumulation for
treating synthetic dyestu wastewater, the microbubble system is a promising process for enhancing both
mass transfer and oxidation of ozone. The total
energy cost should be analyzed further.

Acknowledgement
This research was supported in part by the China Postdoctoral Science Foundation (2005038346).
Appendix A. Supplementary data
Supplementary data associated with this article can
be found, in the online version, at doi:10.1016/
j.chemosphere.2007.03.014.
References
Bader, H., Hoigne, J., 1982. Determination of ozone in water using the
indigo dye method. Ozone Sci. Eng. 4, 169176.
Burns, S.E., Yiacoumi, S., Tsouris, C., 1997. Microbubble generation for
environmental and industrial separations. Sep. Purif. Technol. 11, 221
232.
Chen, Y.H., Chang, C.Y., Su, W.L., Chiu, C.Y., Yu, Y.H., Chiang, P.C.,
Chang, C.F., Shie, J.L., Chiou, C.S., Chiang, S.I., 2005. Ozonation of
CI Reactive Black 5 using rotating packed bed and stirred tank
reactor. J. Chem. Technol. Biotechnol. 80, 6875.
Danckwerts, P.V., 1970. GasLiquid Reactions. McGraw-Hill, New
York.
Fang, X., Mark, G., Sonntag, C., 1996. OH radical formation by
ultrasound in aqueous solutions Part 1: The chemistry underlying the
terephthalate dosimeter. Ultrason. Sonochem. 3, 5763.

Hao, O.J., Kim, H., Chiang, P.C., 2000. Decolorization of wastewater.

Crit. Rev. Env. Sci. Tec. 30, 449505.
Ince, N.H., Tezcanli, G., 2001. Reactive dyestu degradation by combined
sonolysis and ozonation. Dyes Pigments 49, 145153.
Kasprzyk-Hordern, B., Ziolek, M., Nawrocki, J., 2003. Catalytic ozonation and methods of enhancing molecular ozone reactions in water
treatment. Appl. Catal. B-Environ. 46, 639669.
Kodama, Y., Kakugawa, A., Takahashi, T., Kawashima, H., 2000.
Experimental study on microbubbles and their applicability to ships
for skin friction reduction. Int. J. Heat Fluid Fl. 21, 582588.
Mayank, S., Locke, B.R., 2006. Quantication of hydroxyl radicals
produced in aqueous phase pulsed electrical discharge reactors. Ind.
Eng. Chem. Res. 45, 58195825.
Mitani, M.M., Keller, A.A., Sandall, O.C., Rinker, R.G., 2005. Mass
transfer of ozone using a microporous diuser reactor system. Ozone
Sci. Eng. 27, 4551.
Pires, M., Carvalho, L.R.F., 1998. An artifact in air carbonyls sampling
using C-18 DNPH-coated cartridge. Anal. Chim. Acta 367, 223231.
Qiu, Y.Q., Kuo, C.H., Zappi, M.E., 2001. Performance and simulation of
ozone absorption and reactions in a stirred-tank reactor. Environ. Sci.
Technol. 35, 209215.
Rosal, R., Rodriguez, A., Zerhouni, M., 2006. Enhancement of gasliquid
mass transfer during the unsteady-state catalytic decomposition of
ozone in water. Appl. Catal. A-General 305, 169175.
Ruppert, G., Bauer, R., Heisler, G., 1994. UV-O3, UV-H2O2, UV-TiO2
and the photo-fenton reaction comparison of advanced oxidation
processes for waste-water treatment. Chemosphere 28, 14471454.
Serizawa, A., Inui, T., Yahiro, T., Kawara, Z., 2003. Laminarition of
micro-bubble containing milky bubbly ow in a pipe. In: Third
EuropeanJapanese Two-Phase Flow Group Meeting, Certosa di
Pontignano, pp. 2127.
Shin, W.T., Mirmiran, A., Yiacoumi, S., Tsouris, C., 1999. Ozonation
using microbubbles formed by electric elds. Sep. Purif. Technol. 15,
271282.
Shu, H.Y., Huang, C.R., 1995. Degradation of commercial azo dyes in
water using ozonation and UV enhanced ozonation process. Chemosphere 31, 38133825.
Slavinskaya, G.V., Selemenev, V.F., 2003. Ozonation of fulvic acids of
natural waters in aqueous solutions. Russian J. Appl. Chem. 76, 1472
1475.
Soetanto, K., Chan, M., 2000. Fundamental studies on contrast images
from dierent-sized microbubbles: Analytical and experimental studies. Ultrasound Med. Biol. 26, 8191.
Vandevivere, P.C., Bianchi, R., Verstraete, W., 1998. Treatment and reuse
of wastewater from the textile wet-processing industry: Review of
emerging technologies. J. Chem. Technol. Biotechnol. 72, 289302.
Walker, A.B., Tsouris, C., DePaoli, D.W., Klasson, K.T., 2001. Ozonation of soluble organics in aqueous solutions using microbubbles.
Ozone Sci. Eng. 23, 7787.
Wang, C.X., Yediler, A., Lienert, D., Wang, Z.J., Kettrup, A., 2003.
Ozonation of an azo dye CI Remazol Black 5 and toxicological
assessment of its oxidation products. Chemosphere 52, 12251232.
Wu, J.N., Wang, T.W., 2001. Ozonation of aqueous azo dye in a semibatch reactor. Water Res. 35, 10931099.
Zhou, H.D., Smith, D.W., 1997. Process parameter development for
ozonation of kraft pulp mill euents. Water Sci. Technol. 35, 251259.
Zhou, H.D., Smith, D.W., 2000. Ozone mass transfer in water and
wastewater treatment: Experimental observations using a 2D laser
particle dynamics analyzer. Water Res. 34, 909921.